Skip to content

Instantly share code, notes, and snippets.

View wolfram77's full-sized avatar

Subhajit Sahu wolfram77

78 followers · 698 following
View GitHub Profile
@wolfram77
wolfram77 / notes-fast-shared-memory-algorithms-for-computing-the-minimum-spanning-forest-of-sparse-graphs.md
Last active September 16, 2023 19:57
Fast shared-memory algorithms for computing the minimum spanning forest of sparse graphs : NOTES

See below.

@wolfram77
wolfram77 / notes-fast-and-memory-efficient-minimum-spanning-tree-on-the-gpu.md
Last active September 5, 2023 19:10
Fast and Memory-Efficient Minimum Spanning Tree on the GPU : NOTES

Fast and Memory-Efficient Minimum Spanning Tree on the GPU
by Scott Rostrup, Shweta Srivastava, and Kishore Singhal.

@wolfram77
wolfram77 / output-rak-communities-cuda-dynamic--approach-atomic-useold.log
Last active September 5, 2023 17:51
Design of CUDA-based RAK algorithm for community detection : OUTPUT
2023-09-05 21:16:09 OMP_NUM_THREADS=64
2023-09-05 21:16:09 Loading graph /home/graphwork/Data/indochina-2004.mtx ...
2023-09-05 21:16:30 order: 7414866 size: 194109311 [directed] {}
2023-09-05 21:16:36 order: 7414866 size: 341157189 [directed] {} (symmetricize)
{-0.000e+00/+0.000e+00 batchf, 064 threads} -> {0000276.8ms, 0000000.4ms preproc, 0003 iters, 0.904512308 modularity} rakStaticOmpOriginal
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000403.6ms, 0000000.6ms preproc, 0003 iters, 0.903646143 modularity} rakStaticOmp
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000246.8ms, 0000000.6ms preproc, 0006 iters, 0.889725020 modularity} rakStaticCuda
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000163.0ms, 0000000.5ms preproc, 0006 iters, 0.886857124 modularity} rakUseoldStaticCuda
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000140.4ms, 0000000.7ms preproc, 0001 iters, 0.906520933 modularity} rakNaiveDynamicOmp
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000083.2ms, 0000000.6ms preproc, 000
@wolfram77
wolfram77 / output-rak-communities-cuda-dynamic--approach-count-lessatomic.log
Last active September 4, 2023 17:02
Design of CUDA-based RAK algorithm for community detection : OUTPUT
2023-09-04 20:14:42 OMP_NUM_THREADS=64
2023-09-04 20:14:42 Loading graph /home/graphwork/Data/indochina-2004.mtx ...
2023-09-04 20:15:04 order: 7414866 size: 194109311 [directed] {}
2023-09-04 20:15:09 order: 7414866 size: 341157189 [directed] {} (symmetricize)
{-0.000e+00/+0.000e+00 batchf, 064 threads} -> {0000282.3ms, 0000000.4ms preproc, 0003 iters, 0.904451305 modularity} rakStaticOmpOriginal
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000426.3ms, 0000000.8ms preproc, 0003 iters, 0.904793605 modularity} rakStaticOmp
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000265.5ms, 0000000.5ms preproc, 0006 iters, 0.888593117 modularity} rakStaticCuda
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000256.7ms, 0000007.0ms preproc, 0006 iters, 0.889272068 modularity} rakLowatomicStaticCuda
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000136.7ms, 0000004.2ms preproc, 0001 iters, 0.906449489 modularity} rakNaiveDynamicOmp
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000096.8ms, 0000000.6ms preproc,
@wolfram77
wolfram77 / notes-message-from-chairman-and-managing-director-grid-controller-of-india-ltd.md
Last active September 3, 2023 15:00
Message from Chairman and Managing Director-Grid Controller of India Ltd. : NOTES

In this message, Shri K. Sreekant applauds the frontline control room personnel, who are the backbone of GRID-INDIA.

He discusses on the 22.2 GW of wind and solar being scheduled by GRID-INDIA. Due to this, new challenges have emerged, such as Inverter Based Resources (IBRs), and Accuracy of Forecasts. New entities like Battery Energy Storage Systems (BESS) and Green Hydrogen Electrolyzers would soon be part of the electricity grid. To keep pace with these developments, Technical Standards for Grid Connectivity and a robust Compliance Monitoring System is to be put in place. Cyber Security has also emerged as an important area.

Notable achievements in 2022 include Automatic Generation Control (AGC), pilot extension of Security Constrained Economic Despatch (SCED) with ~65 GW capacity under its operation, launch of National Open Access Registry (NOAR), and Green Energy Open Access portal. GRID-INDIA employees wrote and published 23 technical papers. Actions in respect of upgradation of our Ener

@wolfram77
wolfram77 / output-rak-communities-cuda-dynamic--approach-hybrid.log
Last active September 4, 2023 14:42
Design of CUDA-based RAK algorithm for community detection : OUTPUT
2023-09-03 18:41:58 OMP_NUM_THREADS=64
2023-09-03 18:41:58 Loading graph /home/graphwork/Data/indochina-2004.mtx ...
2023-09-03 18:42:20 order: 7414866 size: 194109311 [directed] {}
2023-09-03 18:42:26 order: 7414866 size: 341157189 [directed] {} (symmetricize)
{-0.000e+00/+0.000e+00 batchf, 064 threads} -> {0000257.7ms, 0000000.3ms preproc, 0003 iters, 0.904178602 modularity} rakStaticOmpOriginal
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000387.1ms, 0000000.6ms preproc, 0003 iters, 0.903576723 modularity} rakStaticOmp
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000220.4ms, 0000000.6ms preproc, 0004 iters, 0.900501647 modularity} rakStaticCuda1
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000415.8ms, 0000000.6ms preproc, 0020 iters, 0.904807973 modularity} rakStaticCuda2
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000305.4ms, 0000000.5ms preproc, 0010 iters, 0.903906695 modularity} rakStaticCuda3
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000418.8ms, 0000000.7ms preproc, 0020 iters,
@wolfram77
wolfram77 / output-rak-communities-cuda-dynamic--approach-cross-check-better.log
Last active September 3, 2023 13:10
Design of OpenMP-based RAK algorithm for community detection : OUTPUT
2023-09-03 18:38:31 OMP_NUM_THREADS=64
2023-09-03 18:38:31 Loading graph /home/graphwork/Data/indochina-2004.mtx ...
2023-09-03 18:38:52 order: 7414866 size: 194109311 [directed] {}
2023-09-03 18:38:59 order: 7414866 size: 341157189 [directed] {} (symmetricize)
{-0.000e+00/+0.000e+00 batchf, 064 threads} -> {0000272.8ms, 0000000.8ms preproc, 0003 iters, 0.904668260 modularity} rakStaticOmpOriginal
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000361.0ms, 0000000.7ms preproc, 0003 iters, 0.904419427 modularity} rakStaticOmp
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000213.3ms, 0000000.5ms preproc, 0004 iters, 0.898844303 modularity} rakStaticCuda1
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000412.3ms, 0000000.7ms preproc, 0020 iters, 0.904341266 modularity} rakStaticCuda2
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000292.3ms, 0000000.5ms preproc, 0010 iters, 0.904659465 modularity} rakStaticCuda3
{-5.000e-08/+5.000e-08 batchf, 064 threads} -> {0000402.5ms, 0000000.6ms preproc, 0020 iters,
@wolfram77
wolfram77 / output-pagerank-cuda-dynamic--adjust-switch-degree.log
Last active August 31, 2023 10:09
Design of CUDA-based Dynamic PageRank algorithm for link analysis : OUTPUT
2023-08-23 19:44:14 OMP_NUM_THREADS=64
2023-08-23 19:44:14 Loading graph /home/graphwork/Data/indochina-2004.mtx ...
2023-08-23 19:44:35 order: 7414866 size: 194109311 [directed] {}
2023-08-23 19:44:40 order: 7414866 size: 199021693 [directed] {} (selfLoopAllVertices)
2023-08-23 19:44:42 order: 7414866 size: 199021693 [directed] {} (transposeWithDegree)
{-5.000e-09/+5.000e-09 batchf, 064 threads} -> {0003169.9ms, 074 iter, 4.71e-10 err} pagerankStaticOmp
{-5.000e-09/+5.000e-09 batchf, 064 threads} -> {0000214.0ms, 074 iter, 4.71e-10 err} pagerankStaticCuda16
{-5.000e-09/+5.000e-09 batchf, 064 threads} -> {0000200.9ms, 074 iter, 4.71e-10 err} pagerankStaticCuda32
{-5.000e-09/+5.000e-09 batchf, 064 threads} -> {0000227.1ms, 074 iter, 4.71e-10 err} pagerankStaticCuda64
{-5.000e-09/+5.000e-09 batchf, 064 threads} -> {0000207.9ms, 074 iter, 4.71e-10 err} pagerankStaticCuda128
@wolfram77
wolfram77 / output-pagerank-openmp-dynamic--adjust-prune.log
Last active September 2, 2023 15:45
Design of OpenMP-based Dynamic PageRank algorithm for link analysis : OUTPUT
This file has been truncated, but you can view the full file.
2023-08-23 14:42:03 OMP_NUM_THREADS=32
2023-08-23 14:42:03 Loading graph /home/subhajit.sahu/Data/indochina-2004.mtx ...
2023-08-23 14:42:13 order: 7414866 size: 194109311 [directed] {}
2023-08-23 14:42:18 order: 7414866 size: 199021693 [directed] {} (selfLoopAllVertices)
2023-08-23 14:42:22 order: 7414866 size: 199021693 [directed] {} (transposeWithDegree)
{-5.000e-09/+5.000e-09 batchf, 032 threads, 1e-13 frontier, 0e+00 prune} -> {0003260.8ms, 074 iter, 4.71e-10 err} pagerankStaticOmp
{-5.000e-09/+5.000e-09 batchf, 032 threads, 1e-14 frontier, 0e+00 prune} -> {0003125.0ms, 074 iter, 4.71e-10 err} pagerankStaticOmp
{-5.000e-09/+5.000e-09 batchf, 032 threads, 1e-15 frontier, 0e+00 prune} -> {0003105.4ms, 074 iter, 4.71e-10 err} pagerankStaticOmp
{-5.000e-09/+5.000e-09 batchf, 032 threads, 1e-13 frontier, 1e-10 prune} -> {0003467.9ms, 074 iter, 4.00e-09 err} pagerankPruneStaticOmp
{-5.000e-09/+5.000e-09 batchf, 032 threads, 1e-13 frontier, 1e-11 prune} -> {0003031.0ms, 074 iter, 4.00e-09 err} pagerankPruneStat
@wolfram77
wolfram77 / output-pagerank-cuda-dynamic-frontier.log
Last active August 23, 2023 14:12
Design of CUDA-based Dynamic PageRank algorithm for community detection : OUTPUT
2023-08-23 19:41:18 OMP_NUM_THREADS=64
2023-08-23 19:41:18 Loading graph /home/graphwork/Data/indochina-2004.mtx ...
2023-08-23 19:41:45 order: 7414866 size: 194109311 [directed] {}
2023-08-23 19:41:50 order: 7414866 size: 199021693 [directed] {} (selfLoopAllVertices)
2023-08-23 19:41:52 order: 7414866 size: 199021693 [directed] {} (transposeWithDegree)
{-5.000e-09/+5.000e-09 batchf, 064 threads} -> {0003393.4ms, 074 iter, 4.71e-10 err} pagerankStaticOmp
{-5.000e-09/+5.000e-09 batchf, 064 threads} -> {0000229.5ms, 074 iter, 4.71e-10 err} pagerankStaticCuda16